On the Relationship between the Vocabulary of Bug Reports and Source Code
Authors:
- Laura MorenoLMLaura Moreno
- Wathsala BandaraWBWathsala Bandara
- Sonia HaiducSHSonia Haiduc
- Andrian MarcusAMAndrian Marcus
- 1 September 2013
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 452-455
Abstract
Text retrieval (TR) techniques have been widely used to support concept and bug location. When locating bugs, developers often formulate queries based on the bug descriptions. More than that, a large body of research uses bug descriptions to evaluate bug location techniques using TR. The implicit assumption is that the bug descriptions and the relevant source code files share important words. In this paper, we present an empirical study that explores this conjecture. We found that bug reports share more terms with the patched classes than with the other classes in the system. Furthermore, we found that the class names are more likely to share terms with the bug descriptions than other code locations, while more verbose parts of the code (e.g., comments) will share more words. We also found that the shared terms may be better predictors for bug location than some TR techniques.Keywords
This publication has 9 references indexed in Scilit:
- Structural information based term weighting in text retrieval for feature location2013
- Automatic query reformulations for text retrieval in software engineering2013
- Understanding Android Fragmentation with Topic Analysis of Vendor-Specific Bugs2012
- Clustering Support for Static Concept Location in Source Code2011
- Linguistic analysis of bug report titles with respect to the dimension of bug importance2010
- Mining the coherence of GNOME bug reports with statistical topic models2009
- A Linguistic Analysis of How People Describe Software Problems2006
- An information retrieval approach to concept location in source code2005
- Supporting document and data views of source code2002